Recent studies have begun to identify the component brain structures that participate in mediating human emotional responses. A critical next step is to determine how these structures work together to create different kinds of emotional response. Clues to the functional organization of this system are provided by psychometric and psychophysiologic data suggesting that there are two fundamental dimensions of emotion: valence (pleasantness) and arousal. Unresolved questions include whether valence consists of two independent dimensions (the appetitive and aversive motivational systems) or one bipolar dimensions (the two systems at the extreme of one dimension), whether valence and arousal are independent of one another, and whether the bivariate or bipolar nature of valence varies as a function of arousal level. The study aims to identify the functional anatomy of emotional valence and arousal independent of self-report. The independence of these dimensions will be tested by examining the degree to which the neural activation patterns associated with each dimension are independent, overlapping or reciprocal. These data will therefore directly address the mechanistic questions essential to the functional neuroanatomical model of emotion. Subjects will include 36 men and 18 women ages 18-30, a sampling strategy that will permit both between-sex and within-sex (in males) contrasts and a comparison between the two. PET will be used to measure cerebral blood flow while subjects view picture sets created using the International Affective Picture System. Picture sets will target high arousal pleasant, low arousal pleasant, high arousal unpleasant, low arousal unpleasant and neutral states. A visual fixation condition will also be included to permit between-group comparisons independent of behavioral state. Facial EMG and skin conductance will be obtained as objective measures of valence and arousal to be used as covariates in the analysis of the neuroimaging data. Structural equation modeling of the functional neuroimaging data will then be performed to define the neural networks mediating emotional valence, emotional arousal and their interaction. This study capitalizes on well-characterized methods for eliciting and measuring emotional valence and arousal and a substantial body of preliminary work demonstrating the feasibility and need for the proposed study. This study promises to provide new information about how the neural systems mediating emotion operates in healthy men and women, thereby facilitating future exploration of dysfunction in of this system in psychiatric and systemic medical disorders.